On Sun, Mar 26, 2017 at 11:51 PM, Adam Lackorzynski <adam@os.inf.tu-dresden.de> wrote:
For booting, you'll need a bit of 32bit code, thus the multilib parts
are indeed required.

Thanks for info, nice to learn that!
 
You need to build l4 for L4Linux as L4Linux also requires L4Re.

Hm; so I'm still not exactly clear about how this is structured in the repo and how it builds together. Specifically: I seem to have built the project contained in the l4linux subdir of the l4re-snapshot-2016082114 archive; does it reach from l4linux/ to ../l4re/ and ../kernel/fiasco/ in process, through the makefiles from inside l4linux/ ? It didn't complain about any missing binaries, even though I didn't give it access to any precompiled l4re or fiasco libs before "make".

I seem to have replaced my old kernel with the result, and it went, ummm... suspiciously smooth... am I really running on Fiasco.OC+L4Linux? Or did I just accidentally recompile a regular kernel, and just renamed a few strings in it so that `uname --all` can now wink and say "huh huh, yeah sure, 4.7-l4, nod nod", but in reality I just tricked myself in an overly complicated way?

Is there some means by which I could test/verify that yes indeed, I am now, or am not, on L4 full and proper in my VM? Some L4 API, or some other "one weird trick" I can try, that can only work with L4 (vel Fiasco)? From inside or outside (I'm doing it in Hyper-V)? I've put down the some question in slightly different words on https://superuser.com/q/1193658/12184 if someone would like/prefer that format. If you're curious, the Nix expression I've used can be seen at: https://bitbucket.org/akavel/nixos-hyperv

I'd be very grateful for some more advice!

Thanks & Best Regards,
/Mateusz.